Disability and Rehabilitation

ISSN: (Print) (Online) Journal homepage: https://www.tandfonline.com/loi/idre20

Evaluating validity of the Kids-Balance Evaluation

Systems Test (Kids-BESTest) Clinical Test of
Sensory Integration of Balance (CTSIB) criteria to
categorise stance postural control of ambulant
children with CP

Rosalee M. Dewar, Kylie Tucker, Andrew P. Claus, Wolbert van den Hoorn,
Robert S. Ware & Leanne M. Johnston

To cite this article: Rosalee M. Dewar, Kylie Tucker, Andrew P. Claus, Wolbert van den Hoorn,
Robert S. Ware & Leanne M. Johnston (2021): Evaluating validity of the Kids-Balance Evaluation
Systems Test (Kids-BESTest) Clinical Test of Sensory Integration of Balance (CTSIB) criteria to
categorise stance postural control of ambulant children with CP, Disability and Rehabilitation, DOI:
10.1080/09638288.2021.1887374

To link to this article: https://doi.org/10.1080/09638288.2021.1887374

View supplementary material Published online: 28 Feb 2021.

Submit your article to this journal Article views: 91

View related articles View Crossmark data

Full Terms & Conditions of access and use can be found at

https://www.tandfonline.com/action/journalInformation?journalCode=idre20
DISABILITY AND REHABILITATION
https://doi.org/10.1080/09638288.2021.1887374

ORIGINAL ARTICLE

Evaluating validity of the Kids-Balance Evaluation Systems Test (Kids-BESTest)

Clinical Test of Sensory Integration of Balance (CTSIB) criteria to categorise stance
postural control of ambulant children with CP
Rosalee M. Dewara , Kylie Tuckerb, Andrew P. Clausa, Wolbert van den Hoornc, Robert S. Wared and
Leanne M. Johnstona
a
School of Health and Rehabilitation Sciences, University of Queensland, Brisbane, Australia; bSchool of Biomedical Sciences, University of
Queensland, Brisbane, Australia; cNHMRC Centre of Clinical Research Excellence in Spinal Pain, Injury & Health, Brisbane, Australia; dMenzies
Health Institute Queensland, Griffith University, Nathan, Australia

ABSTRACT ARTICLE HISTORY

Purpose: Evaluate the validity of the Clinical Test of Sensory Integration of Balance (CTSIB) scored using Received 13 May 2020
Kids-Balance Evaluation Systems Test (Kids-BESTest) criteria compared to laboratory measures of pos- Revised 31 January 2021
tural control. Accepted 3 February 2021
Method: Participants were 58 children, 7–18 years, 17 with ambulant cerebral palsy (CP) (GMFCS I–II), and
KEYWORDS
41 typically developing (TD). Postural control in standing was assessed using CTSIB items firm and foam Cerebral palsy; Kids-BESTest;
surfaces, eyes open (EO) then closed (EC). Face validity was evaluated comparing clinical Kids-BESTest CTSIB; postural balance;
scores between groups. Correlating force plate centre-of-pressure (CoP) data and clinical scores allowed school-aged population;
evaluation of concurrent and content validity. validation studies
Results: Face validity: TD children scored higher for all CTSIB conditions when compared to children with
CP. Concurrent validity: the agreement between clinical and CoP derived scores was poor to excellent
(Firm-EO ¼ 76%, Firm-EC ¼ 76%, Foam-EO ¼ 59%, Foam-EC ¼ 94%). Clinical scores of “2-unstable” and
“3-stable” were not distinguished reliably by force plate measures. Content validity: significant correlations
were found between clinical scores and CoP data for the two intermediate conditions (Firm-EC: rs 0.40
to 0.72; Foam-EO: rs 0.12 to 0.50), but not the easier (Firm-EO: rs 0.41 to 0.36) or harder condi-
tions (Foam-EC: rs 0.25 to 0.27).
Conclusion: Face validity of Kids-BESTest CTSIB criteria was supported. Content and concurrent validity
were partially supported. Improved Kids-BESTest scoring terms were recommended to describe postural
characteristics of “2-unstable.”

ä IMPLICATIONS FOR REHABILITATION

 Face validity of the Kids-BESTest criteria for the CTSIB was confirmed.
 The Kids-BESTest criteria for the CTSIB can identify children with atypical postural control.
 Concurrent validity and content validity were partially supported, since children with CP resorted to a
range of different balance strategies when “unstable.”
 To improve CTSIB Kids-BESTest criteria, new terms were recommended to better describe postural
characteristics of “2-unstable.”

Introduction Clinical Test of Sensory Integration of Balance (CTSIB) was devel-

Postural control impairment has a significant impact on activity
and participation for children with cerebral palsy (CP) however
quantification of the validity of assessments for this core deficit is
lacking [1]. Sensory system dysfunction may contribute to poor
control of orientation (alignment of body segments) and/or bal-
ance (maintaining center-of-mass [CoM] within the base-of-sup-
port [BoS]) during standing in children with CP [1–3]. A recent
systematic review has identified a number of clinical assessments
of postural control available to assess aspects of standing orienta-
tion and balance for children with CP, however, these assessments
lack validity data in comparison to laboratory measures [4]. The
oped to assess the relative contribution of visual, vestibular, and
proprioceptive systems towards postural control in quiet stance in
adults [5] and it has been adapted for children [6,7]. It is import-
ant now to determine the validity of clinical ratings for the CTSIB
items for children.
The Kids-Balance Evaluation Systems Test (Kids-BESTest) is the
most comprehensive clinical assessment of postural control for
children with and without motor disorders [6,7]. The Kids-BESTest
“Sensory Orientation” domain includes four CTSIB items (Firm-Eyes
Open (EO), Firm-Eyes Closed (EC), Foam-EO, and Foam-EC), with
scoring criteria reflecting temporal and spatial elements of stance
postural control. Temporal elements quantify the ability to sustain

CONTACT Rosalee M. Dewar rosalee.sheather@uq.net.au School of Health and Rehabilitation Sciences, The University of Queensland, 84a Services Rd, St
Lucia, Brisbane 4072, Australia
Supplemental data for this article can be accessed here.
ß 2021 Informa UK Limited, trading as Taylor & Francis Group
2 R. M. DEWAR ET AL.
stance orientation, that is, either >30 s, 1 < 30 s, or 0 s. Spatial ele-
ments quantify balance, that is, whether the CoM is maintained
within the BoS, in a “stable” or “unstable” manner. These criteria
are valid for rating the performance of adults with balance disor-
ders using the original BESTest [8]. However, the criteria have not
been assessed for validity in children and particularly those
with CP.
Atypical stance postural control has been demonstrated in chil-
dren with CP using force plates [2,3]. Force plates provide data
representing point application of the ground reaction force, or
center-of-pressure (CoP), which reflects postural torques produced
to maintain the CoM within the BoS. Center-of-pressure trajectory
(representing postural sway) is commonly smaller in typically
developing (TD) children than children with motor disorders such
as CP [9]. For example, TD children show less sway with increas-
ing age [10] and when compared to age and gender-matched
children with CP [11,12]. Reduced stance sway following exercise
interventions is also thought to indicate improved postural con-
trol for children with CP [13]. However, sway amplitude alone
does not fully represent postural control in standing. For example,
children with CP with altered sensorimotor control may show var-
iations in CoP path length and/or velocity or use of strategies
such as crouch or co-contraction during quiet stance [14]. A com-
parison of clinical Kids-BESTest CTSIB scores with a variety of CoP
trajectory characteristics is needed to verify the validity of criteria
within this clinical assessment tool.
The aim of this study was to evaluate whether clinical Kids-
BESTest criteria for CTSIB conditions of standing on firm/foam sur-
faces with eyes open/closed show: (i) face validity, by comparing
Kids-BESTest scores between children with and without CP; (ii)
concurrent validity, by comparing CTSIB scores obtained using
Kids-BESTest clinical criteria and CoP derived scores using dur-
ation and range data; and (iii) content validity, by examining rela-
tionships between clinical scores and other CoP trajectory
variables, including velocity, range and area. An additional
exploratory aim was to investigate the range of “unstable” stance
characteristics exhibited by children with CP to assist with clinical
reasoning when assigning quantitative scores.
Methods
Study design and ethics
This was a psychometric study of validity. Data were collected as
part of a larger study on postural control performance in children
with and without CP. Ethical approval was obtained from relevant
Human Research Ethics Committees (NHMRC-EC00179 and
NHMRC-EC00417).
Participants
Participants were children aged seven to 18 years with CP or TD.
Children with CP were recruited as volunteers from a statewide
Table 1. Criteria for clinical scoring (Kids-BESTest) and laboratory scoring (Centre-of-pressure derived scores).
Clinical
Score Kids-BESTest criteria
3 30 s “stable”
2 30 s “unstable”
1 <30 s
0 “unable”
RangeAP/RangeML: range from min to max CoP in antero-posterior or medio-lateral directions; TD: typically developing; Firm-
EO: firm eyes open condition; “unstable”: “to include using a leaning or hip strategy”; “unable”: “cannot stand unassisted.”
CP Register and a statewide clinical CP service. Children were
included in this study if they were: (i) diagnosed with hypertonic-
type CP, (ii) able to stand without support (Gross Motor Function
Classification System [GMFCS] Levels I–III [15]), and (iii) able to fol-
low child-friendly test instructions. Children were excluded if they
had a history of (i) Botulinum Toxin in the last three months, (ii)
orthopaedic or neurological surgery in the previous 12 months, or
(iii) intellectual, behavioural, or medical difficulties that may have
confounded results, or impacted their ability to participate fully in
the assessment.
Typically developing children were recruited as community vol-
unteers. Volunteers were excluded if they were born preterm
(<36 weeks gestation), or if their motor skills were not within the
typical range (percentile rank >15% on the Bruininks–Osteretsky
Test of Motor Proficiency 2nd edition short form [16]). Information
forms were provided to children and their guardians along with a
verbal explanation of the protocol. Prior to participation, guardi-
ans signed a consent form and children signed an assent form.
Procedure
For children with CP classification data was collected regarding
the motor type, GMFCS level, Functional Mobility Scale [17], and
Manual Ability Classification System [18] level. The Kids-BESTest
was administered as per the instructions published by Dewar and
colleagues [7]. The Kids-BEST “Sensory Orientation” domain
includes Item 19, four CTSIB items, as well as Item 20, standing on
an incline board with eyes closed. Only Item 19 was evaluated in
this study. Item 19 involved children performing the four CTSIB
conditions in the following order: (i) Firm-EO, (ii) Firm-EC, (iii)
Foam-EO, and (iv) Foam-EC. Children stood with their feet
together (almost touching) with their hands on their hips. Two tri-
als were performed for each condition. Each trial commenced on
a verbal command “Go” and lasted for a maximum of 30 s.
Between each trial and each condition, participants stepped off
the force plate and walked a short distance (3 m) to a mark on
the floor and back again. Trials were videotaped according to the
Kids-BESTest protocol [7]. Video data were used to support the
examination of content validity.
Outcome measures
CTSIB clinical scoring (Kids-BESTest)
The performance was scored clinically using Kids-BESTest criteria
(from “0 ¼ unable” to “3 ¼ 30 s stable”) (Table 1). Being “unstable”
is defined in the Kids-BESTest criteria “to include using leaning or
hip strategy.” For each condition, children were scored on the first
trial, out of two to reach 30 s, or if 30 s was not reached in either
trial, then the trial with the longest stance duration. Scores for
the four conditions were then summed to give a total score out
of 12.
Laboratory
Centre-of-Pressure derived scoring
30s, with feet in place on the force plate AND
RangeAP and RangeML within ±1.5SD of TD Firm-EO data
30s, with feet in place on the force plate AND
RangeAP and/or RangeML outside ±1.5SD of TD Firm-EO data
0 < 30 s, with feet in place on the force plate
0s, with feet in place on the force plate
 VALIDITY OF THE KIDS-BESTEST CTSIB IN CHILDREN 3

CTSIB laboratory scoring (CoP derived scores) Table 2. Participant characteristics for children with Cerebral Palsy (CP) and typ-
Simultaneous to clinical Kids-BESTest scoring, children stood on a ically developing (TD) children.
force plate (Bertec 4060 series), which provided data for labora- CP Group TD Group
n ¼ 17 n ¼ 41 p-Value
tory validation. Force plate data were digitized (16 bit) at a sam-
pling rate of 1000 samples/s using a Power 1401 data acquisition Male [n (%)] 11 (65%) 21 (51%) 0.40
Age [years] (range) 11.7 ± 2.7 10.9 ± 2.3 0.27
system with Spike2 software (Cambridge Electronic Design (7.9–16.9) (7.8–17.8)
Limited, Cambridge, UK). Force plate data were analyzed with Body mass index [kg/m2] 18.1 ± 4.5 16.8 ± 3.8 0.29
Matlab (R2017a, Matworks Inc., Natick, MA, USA). Data were fil- Height [cm] 146.1 ± 14.4 150 ± 13.9 0.42
Weight [kg] 41.1 ± 15.8 38.9 ± 14.6 0.62
tered using a second-order low-pass bi-directional Butterworth fil-
GMFCS
ter. The cut-off frequency was set at 20 Hz; bi-directional filtering I [n (%)] 11 (65%) N/A
increased the order to four. Then, data were decimated to 100 II [n (%)] 6 (35%)
samples/s for CoP in antero-posterior (AP) and medio-lateral (ML) Manual Ability Classification System (MACS)
I [n (%)] 13 (76%) N/A
vectors to be determined. II [n (%)] 4 (24%)
To examine concurrent validity, CoP data were analysed from Functional Mobility Scale (FMS)
the same trial scored for the Kids-BESTest. Values for RangeAP FMS 5 m score 6a [n (%)] 15 (88%) N/A
and RangeML were calculated from maximum CoP excursion in FMS 50 m score 6a [n (%)] 14 (82%)
FMS 500 m score 6a [n (%)] 13 (76%)
AP and ML directions respectively. Values for TD children for Motor Distribution
RangeAP and RangeML in the baseline condition (Firm-EO) were Hemiplegia [n (%)] 13 (76%) N/A
used as a reference to create categorical scores for “stable/ Diplegia [n (%)] 4 (24%)
unstable” stance for children with and without CP for all condi- Kids-Balance Evaluation Systems Test (Kids-BEST)
Total score /108 75 ± 14.9 99 ± 5.2 <0.01
tions (Table 1). CTSIB score median (range)
To examine content validity, additional CoP characteristics Total /12
were examined in AP and ML directions. These included: All v TD 10 (4–12) 12 (10–12) <0.01
GMFCS I 11 (10–12)
1. Velocity (mm/s); mean of the absolute velocity (VelAP, VelML) GMFCS II 6 (4–11)
2. Peak velocity (mm/s); maximum of the absolute velocity Firm-Eyes Open (/3) 3 (2–3) 3 (3) 0.26
Firm-Eyes Closed (/3) 3 (1–3) 3 (2–3) 0.03
(PeakVelAP, PeakVelML), and
Foam-Eyes Open (/3) 3 (0–3) 3 (2–3) <0.001
3. Area (mm2); total space covered by the CoP trajectory Foam-Eyes Closed (/3) 2 (0–3) 3 (2–3) <0.001
(Area) [19]. Data is reported as Mean ± SD unless otherwise stated.
GMFCS: Gross Motor Function Classification System; CTSIB: Clinical Test of
Possible correlations were then examined between CoP charac- Sensory Interaction on Balance.
teristics and clinical Kids-BESTest scores. a
FMS score for all other children was 5.

Statistical analysis Results

Statistical analyses were performed using Stata statistical software
v13.0 (StataCorp, College Station, TX, USA). Significance was set at
p < 0.05. Group characteristics were compared using Student’s t-
tests, or Chi2 for proportions (Table 2). Face validity was assessed
by comparing clinical CTSIB Kids-BESTest scores (total score and
score for each condition) between children with CP and TD using
the Mann–Whitney test (Table 2).
Concurrent validity was determined by percentage agreement
between clinical CTSIB Kids-BESTest scores as classified according
to Kids-BESTest criteria versus CoP derived scores (Table 1).
Percentage exact agreement was interpreted as: excellent if
>90%, good if >80%, fair if >60% or poor if <60% [6].
Content validity was determined by examining associations
between clinical Kids-BESTest scores (CP/TD) and CoP data for
four CTSIB conditions using mixed-effects linear regression models
(Table 3 and Supplementary Appendix 1). Group and condition
were included as fixed effects, along with a group-by-condition
interaction term. The participant was included as a random effect
to account for the non-independence of the four condition meas-
ures from the same child. Model assumptions were checked
including normality of residuals and homoscedasticity of variance.
To examine whether CoP data were able to reveal additional
information about postural control in quiet standing that could
improve Kids-BESTest criteria for “stable” versus “unstable,”
Spearman rank correlation coefficients (rs) between clinical scores
and CoP were calculated (Table 4). Correlations were interpreted
as: very high ±0.90 to 1.00, high ±0.70 to 0.89, moderate ±0.50 to
0.69, low ±0.30 to 0.49, and negligible ±0.00 to 0.29 [20].
Participants
Families of 62 children responded to the expression of interest,
including 21 children with CP and 41 TD children. Four children
with CP were excluded (n ¼ 1 received lower limb surgery within
12 months, n ¼ 2 had intellectual impairment limiting their partici-
pation, and n ¼ 1 had ataxic CP). Final participants included 58
children aged 7–18 years, 17 with CP (GMFCS I ¼ 11, II ¼ 6;
Diplegia ¼ 4, Hemiplegia ¼ 13; 11 males) and 41 TD children
(Table 2). There were no differences between groups for age, sex
ratio, or body mass index. Of note, although our inclusion criteria
did allow for the recruitment of children up to GMFCS level III, we
had no participant volunteers at this level.
Face validity
Typically developing children performed well on the CTSIB, with
all scoring 2 or 3 points for all conditions according to Kids-
BESTest criteria. The CP group scored lower than the TD group
for every condition, with scores ranging from 1–3 points for firm
conditions and 0–3 points for foam conditions. Statistically, CP
group scores were lower than TD group scores for CTSIB Total
(CP: median 10/12 points; TD: 12/12 points; p  0.001) and three
conditions (Firm-EC, p ¼ 0.03; Foam-EO, p  0.001; and Foam-EC,
p  0.001) but not Firm-EO (p ¼ 0.26) (Table 2). All children with
CP maintained balance during the easiest condition (Firm-EO) for
at least one trial of 30 s. However, some children with CP stood
less than 30 s on both trials for Firm-EC (n ¼ 1), Foam-EO (n ¼ 3),
and Foam-EC (n ¼ 2). Two children with CP were unable to
 4 R. M. DEWAR ET AL.

Table 4. Results of Spearman’s correlations (rs) between clinical Kids-BESTest

Mean diff., 95%CI

The mean difference, including the 95% confident interval between groups, is shown for each variable and condition. Data and comparisons are based on all children in a group completing a given condition, unless

In Firm-EC and Foam-EO, all children with CP were able to complete condition but 15/17 able to complete 30 s; bin Foam-EC, 15/17 children with CP were able to attempt the condition and 13/17 completed 30 s;
11.0, 122.2
scores and CoP data for four CTSIB conditions for children with CP, and one

225, 996
0.8, 12.7

1.2, 13.9

4.3, 13.2

6.9, 10.4
5.4, 117.6
condition (Foam-EC) for TD children.

61.5

55.6

386
6.0

6.4

4.4

1.7
Foam-EC
Firm-EO Firm-EC Foam-EO
CP CP CP CP TD
Foam Eyes Closed

Centre-of-Pressure (n ¼ 17) (n ¼ 17) (n ¼ 17) (n ¼ 15a) (n ¼ 41)

0.41 0.62 0.50 0.27 0.62
(163.4, 335.4)

(177.5, 274.9)
VelAP (mm/s)

(1977, 4690)
(24.1, 48.1)

(28.6, 45.4)

(55.5, 97.9)

(53.6, 76.5)
p-Value 0.10 <0.01 0.04 0.27 <0.01
243.9

236.4

0–30
3038
38.6

37.8

65.9

71.6
CPb

0.41 0.42 0.36 0.22

24
VelML (mm/s) 0.42
p-Value 0.10 0.09 0.15 0.39 <0.01
Peak VelAP (mm/s) 0.41 0.51 0.36 0.09 0.48
p-Value 0.10 0.04 0.16 0.72 <0.01
0.36 0.42 0.42 0.06 0.49
(162.7, 269.6)

(164.2, 258.2)

(2050, 3669)
Peak VelML (mm/s)
(27.5, 43.4)

(28.1, 39.7)

(54.9, 78.2)

(53.2, 79.8) p-Value 0.10 0.09 0.10 0.83 <0.01

30–30
204.4

209.5

2441
32.8

34.0

61.4

63.9
TD

30
RangeAP (mm) 0.41 0.72 0.13 0 0.59
p-Value 0.10 <0.01 0.62 0.85 <0.01
Table 3. Median (interquartile range) of CoP data for the four Kids-BESTest CTSIB items for children with cerebral palsy (n ¼ 17) and typically developing children (n ¼ 41).

RangeML (mm) 0.41 0.4 0.17 0.25 0.44

p-Value 0.10 0.12 0.52 0.33 <0.01
Mean diff., 95%CI

Area (mm2) 0.41 0.54 0.42 0.2 0.51

p-Value <0.01
11.5 3.2,

0.10 0.02 0.09 0.28

465, 1647
6.1, 114.2

0.4, 129.1
5.7, 18.7

3.8, 18.5

3.0, 19.9

1056
12.2

11.1

11.4
60.2

64.7

19.8

Remaining data for Firm-EO, Firm-EC and Foam-EO for TD children is not pro-
vided because a ceiling effect in clinical scores produced no significant associa-
tions with CoP data.
EO: eyes open; EC: eyes closed; CP: cerebral palsy; TD: typically developing; Vel:
Foam Eyes Open

velocity; AP: antero-posterior; ML: medio-lateral.

(106.1, 190.3)

(110.3, 167.1)

n ¼ 15 in this condition as 2 children with CP were unable to stand on foam

(1068, 2753)

a
(18.7, 26.7)

(19.4, 29.8)

(36.0, 61.6)

(39.0, 60.5)

with eyes closed.

142.3

135.7

1–30
1657
21.1

23.1

45.7

47.7
a

26
CP

maintain standing balance during the Foam-EC condition for

any duration.
(79.1, 119.3)

(85.8, 127.7)

(588, 1213)
(12.2, 17.3)

(13.8, 20.0)

(31.5, 44.2)

(29.1, 44.4)

30–30
100.3

107.4
15.3

17.2

36.3

35.9

818
TD

30

Concurrent validity: % agreement

Agreement between clinically rated and CoP derived scores for
Mean diff., 95%CI

the CTSIB was excellent for the most difficult condition (Foam-EC:
12.5, 120.7

13.4, 142.2
2.2, 10.9

2.2, 14.8

88, 1270
0.4, 15.1

2.3, 18.9

94%), good for both firm conditions (Firm-EO: 76%; Firm-EC: 76%),
77.8 *
66.6

10.6

679
7.8
4.4

6.3

but poor for Foam-EO (59%). When children received clinical

mean duration of the longest trial available for analysis. Assumptions checked for linear regression.

scores of 0 points (“unable”) or 1 point (<30 s), agreement with

Firm Eyes Closed

CoP derived scores were 100%. In contrast, almost all occasions of

disagreement (94%), occurred when children received a clinically
(86.1, 136.0)

(70.4, 148.8)

(521, 1546)
(12.8, 18.0)

(12.5, 16.8)

(27.8, 43.5)

(29.3, 59.4)

rated score of 3 points (30 s, “stable”) but a CoP derived a score

103.0

4–30
14.8

13.8

95.4

35.6

36.9

719
a

28
CP

of 2 points (30 s “unstable”: due to CoP range being ±1.5 SD out-

side of the TD reference data).
(78.0, 131.4)

(90.1, 122.0)

(587, 1135)
(12.0, 17.2)

(11.4, 17.8)

(27.9, 39.3)

(30.7, 43.0)

30–30
101.5
14.9

14.3

95.8

32.8

38.2

937

Content validity
TD

30

Children with CP demonstrated higher CoP velocity, range, and

area than TD children however the CoP variable that was signifi-
Mean diff., 95%CI

cantly different depended on the CTSIB condition. For the easiest

29.9, 78.2

36.3, 92.5

200, 982
4.5, 10.2

5.7, 11.3
3.7, 9.4

0.4, 18.0

condition (Firm-EO) children with CP showed a greater RangeML

9.7
24.1

28.1

391
2.9

2.9

2.8

compared to TD children (p < 0.05). For the two intermediate con-

ditions, children with CP showed greater values for all CoP varia-
bles compared to TD children (all p < 0.05), except for Firm-EC,
Firm Eyes Open

RangeAP, and VelAP. For the most difficult (Foam-EC) condition

(63.5, 104.7)
(10.2, 11.0)

(56.7, 96.1)

(22.3, 33.2)

(366, 958)
(9.7, 12.1)

children with CP showed Peak VelAP compared to TD chil-

30–30
(17.7)
10.7

10.8

75.9

75.2

28.1

29.3

631
CP

30

dren (p < 0.05).

When performance on the four CTSIB conditions was com-
pared within a group (Supplementary Appendix 1), TD children
(54.9, 98.2)

(52.3, 82.3)

(20.6, 29.1)

(22.0, 31.7)

(319, 618)
(7.6, 11.4)

(8.0, 11.4)

showed a greater CoP RangeAP in the two most difficult condi-

30–30

p < 0.05; p < 0.01.

76.0

69.0

24.3

27.0

501
9.7

9.4
TD

30

tions (Foam-EO/EC, p < 0.001) compared to the two easiest condi-

tions (Firm-EO/EC, p < 0.001), but performance did not differ
otherwise stated.

between the two intermediate conditions (Firm-EC/Foam-EO,

Timec (sec)
Peak VelML
Peak VelAP

p ¼ 0.18). In contrast, for the CP group, CoP RangeAP increased

RangeML
RangeAP
(mm/s)

(mm/s)

(mm/s)

(mm/s)

with each level of difficulty (Firm-EO/EC, p ¼ 0.001; Firm-EC/Foam-

(mm2)

Range
(mm)

(mm)
VelML
VelAP

Area

EO, p ¼ 0.02; Foam-EO/EC, p  0.001 (Figure 1).

a
c

Figure 1. Scatter plots depicting individual participant values for RangeAP (y-axes) and RangeML (x-axes) for children with typical development (TD, black squares) and
children with cerebral palsy (CP, open circles) for the four Kids-BESTest CTSIB conditions of (A) Firm eyes open (Firm-EO), (B) Firm eyes closed (Firm-EC), (C) Foam-EO,
and (D) Foam-EC. In each diagram, the marked box represents baseline condition (Firm-EO) reference data, that is, 1.5SD from the mean of data for TD children for
RangeAP and RangeML when assessed under (Firm-EO).
Relationship between clinical and laboratory data
When comparing clinical Kids-BESTest scores with CoP data for TD
children, negligible associations were seen for the three easiest
conditions because almost all TD children received a clinical score
of 3 points. In contrast, in the Foam-EC condition, TD children
showed lower clinical scores, which were associated with greater
CoP velocity, excursion, and area (Table 4).
For children with CP, clinical Kids-BESTest scores were signifi-
cantly correlated with CoP data for the two intermediate condi-
tions, Firm-EC and Foam-EO, but not other conditions (Table 4).
For Firm-EC, higher clinical scores (0 points n ¼ 0, 1 point n ¼ 2, 2
points n ¼ 2, 3 points n ¼ 13) were correlated with lower CoP vel-
ocity (VelAP rs ¼ 0.62, p ¼ 0.04; PeakVelAP rs ¼ 0.51, p ¼ 0.04)
and smaller CoP excursion (RangeAP rs ¼ 0.72, p < 0.01; Area rs
¼ 0.54, p ¼ 0.02). For Foam-EO, higher clinical scores (0 points
n ¼ 1, 1 point n ¼ 2, 2 points n ¼ 5, 3 points n ¼ 9) were correlated
only with lower CoP velocity (VelAP rs ¼ 0.50, p ¼ 0.04).
In the most difficult condition, Foam-EC (0 points n ¼ 2, 1 point
n ¼ 2, 2 points n ¼ 11, 3 points n ¼ 2), visual inspection of CoP tra-
jectory plots showed that children with CP demonstrated a range
of atypical control strategies indicative of experiencing difficulty
with maintaining stance, including loss of control (large amplitude
with shortened duration Figure 2, F); asymmetrical leaning (asym-
metrical trajectory Figure 2, N), or overly constrained control
(atypically small trajectory [<1.5SD from reference data] Figure 2,
C). Strategies used by children with CP were further examined
using Kids-BESTest video footage, to confirm five common
VALIDITY OF THE KIDS-BESTEST CTSIB IN CHILDREN 5
strategies: (A) aligned – using an effective ankle strategy, (N) not
aligned – leaning laterally (towards one side) or forward (using a
hip strategy); (C) constraint – using sustained crouch; (V) variable
– alternating between strategies, for example, ankle and/or hip
and/or crouch; or (F) step or fall (Figure 2).
Discussion
This validity study for the CTSIB using Kids-BESTest clinical criteria
for children with and without CP confirmed face validity and par-
tially supported concurrent and content validity. Face validity was
demonstrated as children with CP scored lower Kids-BESTest
scores than TD children. Concurrent validity was partially demon-
strated by good to excellent agreement between clinical and CoP
derived scores for all stance conditions, except Foam-EO, which
had a poor agreement between scores. Content validity was par-
tially supported, since clinical scores correlated well with CoP vari-
ables for the two intermediate conditions, Firm-EC and Foam-EO,
but not with Firm-EO or Foam-EC. Data showed a ceiling effect in
some items for TD children and that children with CP use a var-
iety of atypical balance strategies when struggling with standing
balance. In response to the variety of balance strategies observed,
additional Kids-BESTest criteria descriptors are proposed to
describe variations in “unstable” postural control when performing
the CTSIB.
Face validity was demonstrated for Kids-BESTest Item #19 crite-
ria for children with CP, as they obtained lower CTSIB Total and
6 R. M. DEWAR ET AL.

Figure 2. Example data for children performing the most difficult Foam-EC condition, incorporating images from Kids-BESTest video footage with associated CoP
traces. This data illustrates the five common balance strategies observed for children with CP, including: F ¼ step or fall (large amplitude with short duration); N ¼
“not aligned” (leaning laterally [towards one side – depicted], or forward [using a hip strategy]); C ¼ constraint strategy (use of sustained crouch); V ¼ variable strategy
(alternating between strategies ankle and/or hip strategy and/or crouch); and, A ¼ “aligned” (using effective ankle strategy).

Item (condition) scores than TD children. This shows that overall,
the Kids-BESTest criteria can identify atypical postural control, at
least at a screening level. More information on atypical balance
strategies may be needed to guide treatment as outlined below.
For concurrent validity, the level of agreement between clinic-
ally rated and CoP derived scores varied according to CTSIB con-
dition difficulty. The agreement was highest for the most difficult
condition (Foam-EC) because fewer children completed 30 s of
the task, automatically scoring 1 point (<30 s) or 0 points
(“unable”) irrespective of the atypical balance strategy used. The
agreement was slightly lower for firm conditions. In both firm
conditions, all TD children and most children with CP showed
minimal sway (scoring 3 points clinically), which was easy to
detect using CoP data. However, some children with CP scored 3
points clinically (“stable,” i.e., not moving), but only 2 points
based on CoP data because they used an atypical crouch strategy.
Crouch stance produced an atypically small CoP trajectory
(<1.5SD from the reference data mean), but performance could
not be scored down against the clinical criteria for “unstable,”
which only included “leaning or hip strategy.” The agreement was
lowest for the Foam-EC condition due to a combination of more
frequent use of a crouch strategy (incorrect clinical rating), or
asymmetrical leaning to the hemiplegic side (incorrect CoP
derived rating due to normal CoP range). This again highlighted
the difficulty in applying Kids-BESTest criteria for “unstable” for
children with CP.
A previous examination of Paediatric-CTSIB criteria using
pooled data from children with TD or CP, reported excellent reli-
ability for head sway amplitude when comparing clinical ratings
versus motion analysis scores [21]. This agreement is expected
given the ease of observing head sway clinically. However, head
motion alone would not have identified all instances of atypical
stance control. The study authors confirmed this, suggesting the
measure had only “acceptable levels to detect immature move-
ment strategies” when attempting to maintain a stance.
Content validity analysis revealed that the CTSIB provides an
incremental challenge for children with CP however this popula-
tion uses a range of stance control strategies under postural chal-
lenge that is not distinguished by the current criteria. Although
moderate correlations have been shown between another
method of CTSIB clinical scoring and CoP data in a study involv-
ing children TD 7–12 years [22], this study for children with CP
used Kids-BESTest ratings. For children with CP, the easiest condi-
tion (Firm-EO) showed limited relationships with other CoP varia-
bles due to a ceiling effect in the clinical score (all TD and 16/17
children with CP scoring 3 points). Clinically, this suggests that
Firm-EO is a good reference condition, but it is unlikely to dis-
criminate atypical performance when used alone for ambulant
children with CP. For the most difficult condition (Foam-EC) chil-
dren with CP used multiple compensatory strategies, which
resulted in negligible correlations with CoP data, and emphasized
the need for a broader range of “unstable” descriptors. The

strongest correlations between clinical ratings and CoP data for
children with CP occurred in the two intermediate conditions.
Differences between children in clinical ratings were highest for
these conditions, which led to strong correlations with CoP data
for velocity (Firm-EC and Foam-EO) or CoP excursion (Firm-EC).
Moderate (Firm-EC) or poor (Foam-EO) agreement between clin-
ical Kids-BESTest ratings and corresponding data for CoP range
emphasizes the need to expand Kids-BEStest descriptors to reflect
atypical balance strategies.
Children with CP in our study demonstrated a range of atyp-
ical balance strategies when struggling to maintain stance. This
finding is supported by previous studies of stance postural control
in children with CP [14,23]. We also found that clinical Kids-
BEStest criteria of “unstable” did not clearly encompass some
strategies (e.g., crouch, or variable strategies), and that CoP range
data alone could not always distinguish between strategies with-
out clinical video review (e.g., crouch versus “stable” stance). We
therefore propose prompts to describe five common strategies to
assist with a scoring of the CTSIB (Supplementary Appendix 2).
Strengths, limitations, and future directions for research
This study documents significant new validity data for the CTSIB,
using the Kids-BESTest clinical criteria, for children with and with-
out CP. New prompts have been developed to describe character-
istics of instability and identified differences between children
with and without CP. Future research may be directed to (i) char-
acterise atypical stance strategies used by children with varying
CP severity, motor distribution, motor type, and age; (ii) explore
the validity of other tasks that are part of the Kids-BESTest, and
(iii) include additional outcome measures for postural control
(e.g., kinematics), to augment the information we have reported
from force plate data in standing.
Conclusion
The results of this study support the face validity of the clinical
Kids-BESTest CTSIB criteria to identify atypical postural control in
children with CP. Concurrent validity and content validity were
partially supported since children with CP resorted to a range of
different balance strategies when “unstable,” not just “leaning or
hip strategy” as currently available in the Kids-BEStest. To improve
Kids-BESTest scoring, new terms were recommended to better
describe postural characteristics of “2-unstable.”
Disclosure statement
No potential conflict of interest was reported by the author(s).
Funding
This research was financially supported by the Research
Foundation, Cerebral Palsy Alliance (grant no. PG4114) through
the Children’s Motor Control Research Collaboration.
ORCID
Rosalee M. Dewar http://orcid.org/0000-0002-7020-0396
Robert S. Ware http://orcid.org/0000-0002-6129-6736
Leanne M. Johnston http://orcid.org/0000-0001-6779-4595
VALIDITY OF THE KIDS-BESTEST CTSIB IN CHILDREN 7
References
[1] Dewar R, Claus AP, Tucker K, et al. Perspectives on postural
control dysfunction to inform future research: a Delphi
Study for children with cerebral palsy. Arch Phys Med
Rehabil. 2017;98(3):463–479.
[2] Nashner L, Shumway-Cook A, Marin O. Stance posture con-
trol in select groups of children with cerebral palsy: deficits
in sensory organization and muscular coordination. Exp
Brain Res. 1983;49(3):393–409.
[3] Cherng R-J, Su F-C, Chen J-JJ, et al. Performance of static
standing balance in children with spastic diplegic cerebral
palsy under altered sensory environments. Am J Phys Med
Rehabil. 1999;78(4):336–343.
[4] Saether R, Helbostad JL, Riphagen II, et al. Clinical tools to
assess balance in children and adults with cerebral palsy: a
systematic review. Dev Med Child Neurol. 2013;55(11):
988–999.
[5] Shumway-Cook A, Horak FB. Assessing the influence of
sensory interaction of balance. Suggestion from the field.
Phys Ther. 1986;66(10):1548–1550.
[6] Dewar R, Claus AP, Tucker K, et al. Reproducibility of the
Kids-BESTest and the Kids-Mini-BESTest for children with
cerebral palsy. Arch Phys Med Rehabil. 2019;100(4):
695–702.
[7] Dewar R, Claus AP, Tucker K, et al. Reproducibility of the
Balance Evaluation Systems Test (BESTest) and the Mini-
BESTest in school-aged children. Gait Posture. 2017;55:
68–74.
[8] Horak FB, Wrisley DM, Frank J. The balance evaluation sys-
tems test (BESTest) to differentiate balance deficits. Phys
Ther. 2009;89(5):484–498.
[9] Pavao SL, dos Santos AN, Woollacott MH, et al. Assessment
of postural control in children with cerebral palsy: a review.
Res Dev Disabil. 2013;34(5):1367–1375.
[10] Verbecque E, Vereeck L, Hallemans A. Postural sway in chil-
dren: a literature review. Gait Posture. 2016;49:402–410.
[11] Lopes GHR, David ACd. Posturography in the analysis of
postural control in children with cerebral palsy: a literature
review. Fisioter Pesqui. 2013;20(1):97–102.
[12] Donker SF, Ledebt A, Roerdink M, et al. Children with cere-
bral palsy exhibit greater and more regular postural sway
than typically developing children. Exp Brain Res. 2008;
184(3):363–370.
[13] Dewar R, Love S, Johnston LM. Exercise interventions
improve postural control in children with cerebral palsy: a
systematic review. Dev Med Child Neurol. 2015;57(6):
504–520.
[14] Burtner P, Qualls C, Woollacott M. Muscle activation charac-
teristics of stance balance control in children with spastic
cerebral palsy. Gait Posture. 1998;8(3):163–174.
[15] Palisano RJ, Rosenbaum P, Bartlett D, et al. Content validity
of the expanded and revised Gross Motor Function
Classification System. Dev Med Child Neurol. 2008;50(10):
744–750.
[16] Deitz JC, Kartin D, Kopp K. Review of the Bruininks-
Oseretsky Test of Motor Proficiency, Second Edition (BOT-
2). Phys Occup Ther Pediatr. 2007;27(4):87–102.
[17] Harvey A, Robin J, Morris ME, et al. A systematic review of
measures of activity limitation for children with cerebral
palsy. Dev Med Child Neurol. 2008;50(3):190–198.
[18] Eliasson A-C, Krumlinde-Sundholm L, Ro €sblad B, et al. The
Manual Ability Classification System (MACS) for children
with cerebral palsy: scale development and evidence of
8 R. M. DEWAR ET AL.
validity and reliability. Dev Med Child Neurol. 2006;48(7):
549–554.
[19] Schubert P, Kirchner M. Ellipse area calculations and their
applicability in posturography. Gait Posture. 2014;39(1):
518–522.
[20] Mukaka M. A guide to appropriate use of Correlation coef-
ficient in medical research. Malawi Med J. 2012;24(3):69–71.
[21] Lekskulchai R, Kadli S. Concurrent Validity of the Pediatric
Clinical Test of Sensory Interaction for balance to quantify
postural sway and movement strategies of children aged
7-12 years. J Med Assoc Thai. 2015;98(5):S36–S41.
[22] Pandian TJS, Ukamath S, Jetley N, et al. Clinical test of sen-
sory interaction in balance (CTSIB): concurrent validity
study in healthy Indian children. J Pediatr Neurol. 2011;
9(03):311–318.
[23] Hay L, Redon C. Feedforward versus feedback control in
children and adults subjected to a postural disturbance.
Exp Brain Res. 1999;125(2):153–162.